An activity-based probe for high-throughput measurements of triacylglycerol lipases

Article abstract of DOI:10.1016/j.ab.2011.03.008

Modulating the activity of lipases involved in the metabolism of plasma lipoproteins is an attractive approach for developing lipid raising/lowering therapies to treat cardiovascular disease. Identifying small molecule inhibitors for these membrane-active enzymes, however, is complicated by difficulties associated with measuring lipase activity and inhibition at the water-membrane interface; substrate and compound dynamics at the particle interface have the potential to confound data interpretation. Here, we describe a novel ELISA-based lipase activity assay that employs as "bait" a biotinylated active-site probe that irreversibly binds to the catalytic active-site serine of members of the triacylglycerol lipase family (hepatic lipase, lipoprotein lipase, and endothelial lipase) in solution with high affinity. Detection of "captured" (probe-enzyme) complexes on streptavidin-coated plates using labeled secondary antibodies to specific primary antibodies offers several advantages over conventional assays, including the ability to eliminate enzyme-particle and compound-particle effects; specifically measure lipase activity in complex mixtures in vitro; preferentially identify active-site-directed inhibitors; and distinguish between reversible and irreversible inhibitors through a simple assay modification. Using EL as an exemplar, we demonstrate the versatility of this assay both for high-throughput screening and for compound mechanism-of-action studies.

Full text of DOI:10.1016/j.ab.2011.03.008

Analytical Biochemistry 414 (2011) 254–260
Contents lists available at ScienceDirect
Analytical Biochemistry
journal homepage: www.elsevier.com/locate/yabio
An activity-based probe for high-throughput measurements
of triacylglycerol lipases
John Tam a, Martin Henault a, Lianhai Li b, Zhaoyin Wang b, Anthony W. Partridge a, Roman A. Melnyk c,
⇑
a Department of In Vitro Sciences, Merck Frosst Centre for Therapeutic Research, 16711 Trans Canada Highway, Kirkland, Quebec, Canada H9H 3L1
b Department of Medicinal Chemistry, Merck Frosst Centre for Therapeutic Research, 16711 Trans Canada Highway, Kirkland, Quebec, Canada H9H 3L1
c Department of Biology, Merck Frosst Centre for Therapeutic Research, 16711 Trans Canada Highway, Kirkland, Quebec, Canada H9H 3L1
a r t i c l e i n f o
a b s t r a c t
Article history:
Modulating the activity of lipases involved in the metabolism of plasma lipoproteins is an attractive
approach for developing lipid raising/lowering therapies to treat cardiovascular disease. Identifying small
molecule inhibitors for these membrane-active enzymes, however, is complicated by difficulties associ-
ated with measuring lipase activity and inhibition at the water–membrane interface; substrate and com-
pound dynamics at the particle interface have the potential to confound data interpretation. Here, we
describe a novel ELISA-based lipase activity assay that employs as ‘‘bait’’ a biotinylated active-site probe
that irreversibly binds to the catalytic active-site serine of members of the triacylglycerol lipase family
(hepatic lipase, lipoprotein lipase, and endothelial lipase) in solution with high affinity. Detection of ‘‘cap-
tured’’ (probe–enzyme) complexes on streptavidin-coated plates using labeled secondary antibodies to
specific primary antibodies offers several advantages over conventional assays, including the ability to
eliminate enzyme–particle and compound–particle effects; specifically measure lipase activity in com-
plex mixtures in vitro; preferentially identify active-site-directed inhibitors; and distinguish between
reversible and irreversible inhibitors through a simple assay modification. Using EL as an exemplar, we
demonstrate the versatility of this assay both for high-throughput screening and for compound mecha-
nism-of-action studies.
Received 1 February 2011
Received in revised form 3 March 2011
Accepted 5 March 2011
Available online 11 March 2011
Keywords:
Activity-based probe
Active-site ELISA
Reversibility
Occupancy
HDL
Cholesterol
Ó 2011 Elsevier Inc. All rights reserved.
Members of the triacylglycerol lipase family, including endothe-
lial lipase (EL),1 lipoprotein lipase (LPL), and hepatic lipase (HL) are
sn-1 lipases that play a central role in triglyceride and/or phospho-
lipid hydrolysis within plasma lipoproteins. Despite the high se-
quence homology among EL, LPL, and HL [1], they have been
shown to display distinct preferences both for their lipid substrates
and for the particles on which they act: EL predominantly hydrolyzes
phospholipids in HDL particles; LPL predominantly hydrolyzes tri-
glycerides in ApoB-containing particles; and HL readily hydrolyzes
both triglycerides and phospholipids in all classes of lipoproteins
[2]. Extensive knockout and overexpression studies in mice have
confirmed the importance of EL, HL, and LPL mass on the overall lipid
profile [3–7]. Recently, a double EL:HL knockout mouse was con-
structed, both of which demonstrated the complementary effects
of each lipase on the metabolism of HDL and apoB-containing lipo-
proteins, and highlighted the potential importance of developing
EL inhibitors that are exquisitely selective against HL, in addition
to LPL [8].
Assays designed to measure lipase activity and identify inhibi-
tors or activators thereof typically use radiolabeled or fluorescently
labeled lipid substrates emulsified in glycerol and/or non-ionic
detergents such as Triton X-100 [9]. More sophisticated lipase as-
says using synthetic- or native-HDL particles have been developed
recently and are amenable for screening large sample sets in high-
throughput mode [10,11]. While each of these formats offers un-
ique capabilities, the requirement for inclusion of lipid particles
in these assays introduces potential artifacts related to com-
pound–particle interactions, which could dramatically alter the
apparent potency of a particular compound, through either pertur-
bation of the particle/substrate or sequestration of compound
within particles. One approach to circumvent such lipid particle ef-
fects is to measure lipase activity using a soluble substrate, usually
a short alkyl chain linked through an ester linkage to a chromo-
phore that fluoresces on cleavage, which does not require emulsi-
fication. The general utility of the small, soluble substrate
approach, however, is limited by the promiscuity of such sub-
strates to esterases in general, precluding lipase-specific measure-
ments in complex mixtures. Moreover, care must be taken to
⇑
Corresponding author. Address: The Hospital for Sick Children, 555 University
Avenue, Toronto, Ontario, Canada M5G1X8.
E-mail address: roman.melnyk@sickkids.ca (R.A. Melnyk).
Abbreviations used: ABP, activity-based probe; DMSO, dimethyl sulfoxide; EL,
1
endothelial lipase; HL, hepatic lipase; HUVEC, human umbilical vein endothelial cells;
LPL, lipoprotein lipase; PBS, phosphate-buffered saline; SEM, standard error of the
mean; TNF, tumor necrosis factor.
0003-2697/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.ab.2011.03.008

Active-site ELISA for Lipase Activity / J. Tam et al. / Anal. Biochem. 414 (2011) 254–260
255
ensure that substrates remain ‘‘soluble’’, and do not form mixed
Chemical synthesis of the activity-based probe (ABP)
micelles in the presence of certain compounds.
Here, we describe the development of a particle-free, lipase ac-
tive-site occupancy assay using a biotinylated activity-based probe
that forms an irreversible covalent bond with the lipase active-site
catalytic serines through a designed sulfonyl fluoride warhead.
Detection of streptavidin-captured probe:lipase complexes with
The scheme for the synthesis of the ABP is shown below. All re-
agents and solvents used in the synthesis were from a commercial
source and used as received. The newly synthesized compounds
were all characterized by HNMR and FNMR and agreed with their
assigned structures.
antibodies specific to either EL or LPL engenders the ability to dif-
ferentially and specifically measure the activity and inhibition of
different lipases simultaneously, and in complex mixtures.
Through simple modifications of this assay, we further demon-
strate the utility of this approach for critical inhibition mecha-
nism-of-action studies.
The synthesis of biotinylated M352 was achieved in four steps
with commercially available compounds 1 as the staring material.
By refluxing equimolar amounts of 1, 2, and PCl3 in xylene, com-
pound 3 was obtained. Under the cocatalyzation of CuI (0.15 eq)
and PdCl2(PPh3)2 (0.1 eq) and in the presence of 1 eq of K2CO3
as base, the coupling between 3 and 4 led to the formation of com-
pound 5. The treatment of compound 5 with a mixture of TFA/
DCM/H2O at rt for 30 min yielded the corresponding acid 6. In
the presence of HATU (1 eq) and DIPEA (1 eq), equimolar amounts
of compound 6 and compound 7 were coupled to formed the final
biotinylated M352.
Materials and methods
Materials
Compound inhibitors
Plasticware used for cell culture and assays was purchased
from Fisher Scientific (Pittsburgh, PA). Streptavidin Hi-bind black
plates, Superblock buffer, and Quantablue peroxidase substrate
were purchased from Pierce (Rockford, IL). Polyclonal anti-EL
antibodies ab14797 and ab14796 and anti-LPL antibody 21356
were purchased from Abcam (Cambridge, MA), and anti-rabbit
peroxidase antibody RPN4301 was purchased from GE Health-
care (Baie d’Urfe, QC). Human umbilical vein endothelial cells
(HUVEC) and EGM media (containing a final concentration of
2% FBS) were purchased from Lonza (Basel, Switzerland). Expres-
sion plasmids encoding wild-type EL and active-site S169A
mutant EL were prepared by inserting the coding sequence of
the respective genes into the pcDNA3 vector (Invitrogen,
Carlsbad, CA). DNA for transfections was prepared by maxiprep
(Promega, Madison, WI).
The ABP is a biotinylated derivative of a sulfonyl fluoride mole-
cule that covalently and irreversibly binds to the active site of li-
pases (M-352). The proprietary test compounds M-352, M-217,
M-890, M-792, M-973, and M-809 were prepared at Merck. The
pan-lipase inhibitor orlistat [12–14] is from Roche Pharmaceuticals
(San Francisco, CA). Compounds being tested were first serially di-
luted in DMSO before adding to the enzyme reaction mixture.
Cell culture
Human umbilical vein endothelial cells were grown in EGM
media containing FBS until confluency. EL protein expression was
induced by replacing the growth media with serum-free EGM con-
taining 10 ng/mL TNFa (R&D Systems, Minneapolis, MN). The next

256
Active-site ELISA for Lipase Activity / J. Tam et al. / Anal. Biochem. 414 (2011) 254–260
day heparin (Sigma, St. Louis, MO) was added to a final concentra-
tion of 10 U/mL to dissociate EL from the cell surfaces, and the
media were collected and concentrated 250-fold by ultrafiltration
prior to storage at ꢀ20 °C. The concentration of active enzyme
was estimated by measuring the hydrolysis of Bodipy-labeled
phospholipid particles, with 1 unit defined as the amount of en-
zyme able to generate a 10-fold increase in fluorescence after a
20-min reaction time [10].
Alternatively, recombinant EL protein was obtained by transient
transfection of HEK293-T cells with plasmid vectors encoding the
full-length wild-type or S169A active-site mutant EL sequences.
Cells were plated in 225-cm tissue culture flasks and cultured in
DMEM media containing 10% FBS until they reached 90% conflu-
Immunoprecipitation and Western blots
In order to detect the interaction of the ABP with EL by immu-
noblot, EL pretreated with vehicle (DMSO) or 5
60 min was incubated with 100 nM ABP for 60 min before quench-
ing with 1 M nonbiotinylated probe M-352. For each condition,
lM orlistat for
l
sufficient material corresponding to 12 wells of the active-site ELI-
SA was prepared. Streptavidin Sepharose (GE Healthcare) was
added and the reaction was incubated on a rocking platform mixer
for 60 min. The resin was then pelleted by centrifugation, and
washed with assay buffer. The unbound material was concentrated
10-fold using Microcon YM-50 ultrafiltration devices (Millipore,
Billerica, MA) and retained for Western blot analysis. The Sephar-
ose resin was boiled for 5 min with Laemmli loading buffer plus
b-mercaptoethanol (Bio-Rad, Mississauga, ON) to solubilize and re-
lease the ABP:EL complexes, and the samples were loaded on a 4–
20% gradient SDS polyacrylamide gel (Invitrogen). Following elec-
trophoresis, samples were transferred to nitrocellulose using an
iBlot device (Invitrogen), blocked with 5% milk/TBS, and probed
with a 1/3000 dilution of the ab14797 anti-EL antibody. Following
an overnight incubation with the primary antibody, the blot was
washed with TBS/0.1% Tween 20 and incubated with a 1/5000 dilu-
tion of anti-rabbit horseradish peroxidase (GE Healthcare) for
60 min. After the final washes in TBS/Tween 20, chemiluminescent
detection was carried out using Supersignal West Femto substrate
(Pierce) and exposing to Biomax MR film (Kodak, Rochester, NY).
ency. Transfection of 40 lg of plasmid DNA was performed using
Lipofectamine 2000 (Invitrogen) according to the manufacturer’s
instructions, and the media were replaced 6 h post transfection
with Optimem media (Invitrogen). The EL was dissociated and col-
lected 72 h later by the addition of heparin to a final concentration
of 10 U/mL for 30 min. The media were then concentrated 15-fold
by ultrafiltration prior to storage at ꢀ20 °C. The concentration of
active recombinant enzyme was estimated by a comparison with
the activity of a known quantity of HUVEC-derived EL. The concen-
tration of the S169A active-site mutant was normalized to the
wild-type enzyme by measuring Western blot band intensity.
Active-site ELISA
The final assay conditions after optimization were as follows:
Fluorescence-based micellar lipase assay
For each well in a polypropylene 96-well plate, 100
fer (PBS/0.01% TX100) containing 3 nM EL enzyme was mixed with
0.5 L of compound or vehicle (DMSO), and incubated for 60 min.
The ABP was added in a volume of 0.5 L for a final concentration
lL of assay buf-
To test lipase activity and inhibition in a more standard micelle-
based assay, we employed a fluorescence-based micellar lipase as-
say, described by Mitnaul et al. [10]. Briefly, test compounds were
diluted in DMSO, added to 96-well plates containing the lipase en-
zyme in assay buffer (PBS, 10% DMEM, 1.5% glycerol, 0.5% BSA), and
incubated for 30 min prior to adding the phospholipid particles
(Mono-Bodipy-TG or Bis-Bodipy-PC in 0.1% Triton X-100) to a final
l
l
of 50 nM, and incubated for 10–15 min or 120 min for standard
and reversibility conditions, respectively. Next, the reaction was
quenched using an excess (1 lM) of nonbiotinylated probe M-
352 before transferring to a prewashed (PBS/0.1% Tween) Hi-bind
streptavidin multiwell plate. The streptavidin plate was incubated
while shaking for 30 min to allow binding of the enzyme–biotin
probe complexes to the immobilized streptavidin. The plate was
then washed four times with wash buffer (PBS/0.1% Tween) using
a Skatron plate washer (Molecular Devices, Toronto, ON) to remove
unbound enzyme. Bound EL was detected by the combined addi-
tion of a 1/1500 dilution of the ab14797 and ab14796EL antibodies
in PBS/0.1X Superblock, and incubation for 60 min, followed by an-
other plate wash step. Alternatively, the detection of bound LPL
was performed using a 1/5000 dilution of ab21356. The sandwich
ELISA was completed by incubating PBS/0.1X Superblock contain-
ing a 1/8000 dilution of anti-rabbit (for anti-EL polyclonal antibod-
ies) or anti-mouse (for anti-LPL monoclonal antibody) horseradish
peroxidase for 60 min. After a final plate wash step, peroxidase
concentration of 20 lM. The increase in fluorescence over time was
read in a plate reader at ex/em of 550/590 nm.
Results and discussion
Design of a lipase-directed activity-based probe
In designing a prototypic lipase-specific ABP, we sought a
molecular scaffold that satisfied the following criteria: straightfor-
ward chemical synthesis; stable in aqueous solution; similar po-
tency against EL, HL, and LPL; and containing a warhead specific
for serine-hydrolases. We focused on compounds with greater li-
gand efficiency containing a sulfonyl fluoride warhead—a widely
studied rapid and irreversible inhibitor of serine hydrolases [15].
M-352 (Fig. 1) emerged as a promising parent scaffold molecule,
based on its overall chemical properties and potency profile. M-
352 potently inhibited EL, HL, and LPL in micellar assays with
low nanomolar potency in a time-dependent manner that was un-
changed by rapid dilution, consistent with the expected irrevers-
ible mode of inhibition (see below). In order to convert M-352
into a prototypic activity-based probe, we incorporated a biotin
moiety distal to the sulfonyl fluoride warhead separated by a flex-
ible linker (Fig. 1). The linker was included to provide sufficient
distance for the biotin to ‘‘snorkel’’ out of the deep lipase active
sites to bind streptavidin unhindered. The fully elaborated ABP re-
tained significant potency against all three lipases when tested in a
standard micellar lipase assay: EL IC50 = 47 nM; HL IC50 = 73 nM;
and LPL IC50 = 73 nM.
activity was measured by adding 100
lL/well Quantablue sub-
strate for 10 min followed by 100 L/well Quantablue stop solu-
l
tion, and the fluorescence signal was measured in a Gemini plate
reader (Molecular Devices) at an ex/em of 325/420 nm. Note that
the entire assay was performed at room temperature.
Assay optimization was performed by varying the concentra-
tion of the ABP, the incubation time, the anti-EL antibodies, and
the detergent concentration of the assay buffer. The 96-well assay
was expanded to 384-well format using Biomek FX robotics (Beck-
man, Mississauga, ON), with assay conditions as described above,
except for an assay volume of 50 lL and the use of 384-well plas-
ticware. The Z0 value was calculated using the equation
Z0 = 1 ꢀ [(3sf + 3sb)/|lb
ꢀ
l
f|],
where
s = standard
deviation,
l
= average, f = background control, and b = positive control. There
was no significant change in assay performance between 0% and 2%
DMSO final concentration.

Active-site ELISA for Lipase Activity / J. Tam et al. / Anal. Biochem. 414 (2011) 254–260
257
Fig.2. Western blot detection of EL–ABP complexes pull-down by streptavidin
beads. (A) Recombinant EL was visualized by immunoblot following incubation
with ABP, capturing to streptavidin Sepharose, and elution with Laemmli sample
buffer. EL-ABP complexes were detected using an EL-specific polyclonal antibody or
an anti-biotin antibody. Where indicated, EL was pretreated with 5 lM orlistat, or
not incubated with ABP. SDS-resistant EL–ABP complexes were detected in the
streptavidin-bound fraction only when incubated with the ABP, and this effect was
abolished by pretreatment with orlistat. The results are representative of three
experiments. (B) The ABP binds exclusively to wt full-length EL. Recombinant wild-
type (wt) or serine 169 to alanine mutant EL was visualized by immunoblot
following incubation with ABP, capturing to streptavidin Sepharose, and elution
with Laemmli sample buffer. EL was detected using an EL-specific polyclonal
Fig.1. Chemical structure of the activity-based probe (ABP) and potency against
triacylglycerol lipases. The ABP was based on a potent pan-selective, irreversible
triacylgycerol lipase inhibitor, M-352.
antibody. Where indicated, the wild-type EL was pretreated with 5 lM orlistat, or
not incubated with ABP. Arrows indicate full-length and cleaved EL. The full-length
Characterization of the ABP
wild-type EL, but not the cleaved EL or active-site mutant EL, was labeled by the
ABP.
As an initial proof-of-principle set of experiments, we first
tested the ability of the ABP to recognize and bind EL in a complex
heterogeneous mixture. Primary HUVEC cells were treated with
the full-length and catalytic domain-only truncated form, which
arises from cleavage immediately downstream of the RNKR site
of the full-length enzyme by endogenous proprotein convertases
[19]. Interestingly, the ABP bound to the full-length form of EL,
and not the truncated catalytic domain, suggesting that the com-
plete enzyme, including the carboxy terminus lipid-binding do-
main is required for substrate/inhibitor engagement. The
apparent absence of enzyme activity in the truncated EL observed
in the present study is in agreement with the findings of Gauster
et al. [20], who reported that the full-length form of EL is cleaved
by endogenous proprotein convertases to a truncated form, leading
to loss of HDL hydrolyzing activity. Cleavage of EL appears to be a
common occurrence in native endothelial cell lines as well [19].
Moreover, cleavage of EL is not an exclusive feature of mammalian
cells; it was noted that expression of recombinant EL in a baculo-
virus/insect cell system yielded a similar sized cleavage product
(A. Auger, unpublished results).
TNF-
a to specifically induce the expression of EL [16,17,10]. Cells
were then treated with heparin to release heparin sulfate-associ-
ated EL into the medium. EL-containing cell supernatants were
then pretreated for 30 min either with DMSO or with the pan-li-
pase inhibitor tetrahydrolipstatin (THL, or orlistat), followed by a
30-min incubation with the ABP, after which the streptavidin Se-
pharose was added to ‘‘capture’’ EL:ABP complexes. After extensive
washing EL was released from streptavidin by boiling the resin in
SDS-Laemmli sample buffer, followed by SDS–PAGE separation
and visualization by immunoblot using both an EL-specific poly-
clonal antibody and an anti-biotin antibody (Fig. 2). As shown in
Fig. 2A, EL was detected in the streptavidin-bound fraction when
incubated with the ABP, and this band was not detected on pre-
treatment with orlistat. The ability of the irreversible lipase inhib-
itor orlistat to prevent binding of the ABP is consistent with their
expected overlapping binding sites (i.e., the active site). On probing
with an anti-biotin antibody, a weak, but reproducible signal was
detected at the expected molecular weight of EL, consistent with
an irreversible covalent ABP-EL interaction. To further confirm that
binding of the ABP to EL was via the expected irreversible trapping
of the ABP to the active-site catalytic serine, we measured the abil-
ity of the ABP to bind to the active-site mutant recombinant S169A
as compared with wild-type EL. The ABP did not label the S169A
active-site mutant of EL (Fig. 2B), confirming both active-site
engagement and previous findings that this construct is catalyti-
cally inactive [18].
Development and optimization of an ABP-based ELISA
We next sought to use the ABP to develop an ELISA-based assay
to quantitatively measure lipase activity and inhibition in a format
that could potentially be used for high-throughput measurements
of activity and inhibition. To determine the optimal probe binding
conditions in the context of a particle-free, ELISA-based binding as-
say, a time course was performed using different concentrations of
the probe with a fixed concentration of EL enzyme (3 nM).
Whereas no signal was observed in the absence of probe, the mag-
Also observed in the loading control lanes (Fig. 2B) was that the
recombinant EL stocks used for these experiments contained both

258
Active-site ELISA for Lipase Activity / J. Tam et al. / Anal. Biochem. 414 (2011) 254–260
Fig.3. ABP time course and titration. EL from HUVEC conditioned media was incubated with 0–250 nM ABP for 0–90 min before quenching with 1
lM M-352. Binding of
biotinylated EL to the streptavidin plate and detection by ELISA was as described under Materials and methods. Bars represent SEM of duplicate wells.
nitude of ELISA fluorescent signal associated with probe-bound EL
increased with increasing probe concentration up to 200 nM, but
began to display diminished signal at a concentration of 250 nM
(Fig. 3). Furthermore, when the probe was tested at concentrations
beyond 150 nM, assay signal amplitude loss became apparent at
the 90-min time point. The diminishing returns at higher probe
concentrations reflect interference by increasing concentrations
of free probe, which begins to compete with probe–enzyme for
binding sites on the streptavidin plate, thereby reducing observed
signal. Based on the data presented in Fig. 3, we used a probe con-
centration and incubation time of 50 nM and 10 min, respectively,
for subsequent experiments. Such conditions were chosen to en-
sure that sufficient assay signal is achieved, while minimizing the
displacement of enzyme-bound compounds over time, and avoid-
ing conditions that lead to loss of signal. Using such optimized con-
ditions produced a >10-fold window above mock-transfected
background that was dose dependently decreased by orlistat
(Fig. 4).
and as a pooled mixture; it was found that pooling the two anti-
bodies provided up to 60% greater signal compared to using either
antibody alone (data not shown). Finally, we measured the effect of
increasing concentrations of Triton X-100 to the assay protocol.
The use of sub-CMC concentrations of non-ionic detergents has
previously been shown to be beneficial to the general performance
of enzyme assays, by reducing the incidence of compound or en-
zyme aggregation [21]. We found that sub-CMC concentrations of
Triton X-100 were tolerated, and in fact improved; addition of
0.001% (v/v) and 0.01% (v/v) Triton X-100 increased assay signal
by 33& and 100%, respectively. As expected, above the CMC of Tri-
ton X-100 (CMC ꢁ0.033%) signal loss was evident: 45% signal loss
at 0.1% Triton X-100. The use of Tween 20 and heparin was also
evaluated, but was not found to significantly improve the assay
signal (data not shown).
Inhibitor titrations and optimization of the ABP-ELISA to 384-well
format
To further improve the sensitivity of the assay, we combined
two commercially available antibodies directed against mutually
exclusive regions of EL. The antibodies raised against the amino-
terminal and carboxy-terminal EL were tested both individually
Initially, the performance of the ABP-ELISA assay was tested in a
dose–response format in 96-well plates. We selected a panel of
structurally diverse inhibitors of EL with a range of potencies, as
measured in standard in vitro micelle-based lipase assays. As
shown in Fig. 5A, test compounds yielded classic dose–response
curves with IC50 values that agreed well with their potencies in
micellar assays (i.e., M-352 was <2-fold shifted and orlistat was
<3-fold shifted from the micellar assay to the EL ELISA). To facili-
tate automated high-throughput screening, the 96-well assay
was miniaturized to a 384-well format and liquid dispensing was
automated using robotics. Enzymes with DMSO were added to rep-
licate wells of a 384-well plate in columns 1, 2, 3, 5, 9, 13, 16, 19,
22, 23, and 24, and compared to the basal signal in the DMSO plus
buffer-only controls in columns 4 and 21 (Fig. 5B). The measured
average fluorescence in the enzyme-containing wells was 7335
units, compared to 1435 average fluorescent units measured in
the control wells, while the Z0 was calculated to be 0.62 (Fig. 5C).
Fig.4. The wild-type but not the catalytically dead S169A mutant was labeled by
the ABP and detected by ELISA. Orlistat was serially diluted in DMSO and added to
3 nM recombinant wild-type sequence (wt) or serine 169 to alanine active-site
mutant (S169A) EL from HEK293 culture media for 60 min. As a negative control,
culture media from mock-transfected cells were used. The ABP was added to a final
Reversibility mode
To develop a format of the assay that could potentially measure
the reversibility of test compounds—a desirable property for a po-
tential small molecule lead—we took advantage of the time-depen-
concentration of 50 nM for 10 min before quenching with
1 lM M-352 and
applying to a 96-well streptavidin plate. Detection of bound EL by ELISA was as
described under Materials and methods. Bars represent SEM of duplicate wells.

Active-site ELISA for Lipase Activity / J. Tam et al. / Anal. Biochem. 414 (2011) 254–260
259
Fig.6. Extending probe incubation time to identify reversible compounds. Recom-
binant EL from HEK293 media was incubated with 50 nM ABP (squares) or no probe
(circles) for 0, 10, 30, 60, 120, 180, and 240 min before quenching with 1 lM M-352
and applying to a streptavidin plate. Detection of bound EL by ELISA was performed
as described under Materials and methods. Arrows indicate the time points chosen
to test compounds under standard and reversibility conditions, respectively. Bars
represent SEM of duplicate wells.
analog of M-352 with a Boronic acid warhead in place of a sulfonyl
fluoride); and a suspected reversible compound (M-809—a close
analog of M-352 with a methyl in place of the warhead). A right-
ward shift in the titration curves for M-809 and M-973 was ob-
served with the extended probe incubation time, consistent with
their expected reversible binding mode (Fig. 7). In contrast, the
measured potency of the nonbiotinylated sulfonyl fluoride M-352
was not affected by the ABP incubation time, thus indicating that
this compound could not be displaced by the probe once bound
to the active site, and is therefore an irreversible compound.
Utility of the ABP against other lipases
To demonstrate that ABP could be used to measure a different
triglyceride lipase, we modified the assay for LPL activity measure-
ments. Orlistat was serially diluted in DMSO and added to diluted
culture media containing 3 nM recombinant EL or LPL. As a nega-
tive control, culture media from mock-transfected cells were used.
The mixture was then incubated with ABP before applying to the
streptavidin plate. When an anti-LPL-specific antibody was used,
bound LPL could be detected, and this signal was dose dependently
attenuated by preincubation with the pan-lipase inhibitor orlistat
(Fig. 8). The LPL-specific antibody did not cross-react with EL or
with proteins from mock-transfected cell culture media.
Fig.5. (A) Titration of compounds by active-site ELISA. Each compound was serially
diluted in DMSO and added to 3 nM EL from HUVEC conditioned media. After
60 min ABP was added to a final concentration of 100 nM and incubated for 15 min.
The reaction was quenched with 1 lM M-352, and added to a streptavidin plate.
The ELISA was performed using anti-EL polyclonal antibodies and detection was by
a fluorogenic peroxidase substrate as described under Materials and methods. Bars
represent SEM of duplicate wells. (B) Adaptation of the active-site ELISA from 96- to
384-well format. Enzyme + DMSO was added to all vertical columns 1, 2, 3, 5, 9, 13,
16, 19, 22, 23, and 24 of a 384-well plate. DMSO alone was added to columns 4 and
21. (C) Assay statistics.
Conclusion
Assays of lipase activity typically employ fluorescent or radioac-
tive lipid-like substrates that are embedded in lipid emulsions.
Though generally robust, when applied to the identification of
inhibitors in high-throughput screens, these assays are often prone
to artifacts arising from signal interference and/or compound–lipid
interactions. Here, we describe the development and optimization
of a particle-free ELISA-based assay that is capable of measuring
active-site engagement of triglyceride lipase family members in
complex mixtures through the use of a designed activity-based
probe. The assay has excellent statistics in high-throughput
screening mode and can be used to distinguish reversible from
irreversible inhibitors. Given the versatility and uniqueness of this
approach, relative to existing assays, we anticipate that this assay
and probe will be very useful for researchers studying lipases, and
in particular for those attempting to identify inhibitors of triglycer-
ide lipase family members.
dent inhibition of the ABP to develop a reversibility assay. We rea-
soned that longer ABP incubation would allow for displacement of
reversibly bound inhibitor–enzyme complexes, leading to greater
probe binding at a given concentration of inhibitor. A reversible
compound would thus appear less potent on increased probe incu-
bation, whereas an irreversible inhibitor would maintain its po-
tency. As seen in Fig. 6, a time course of enzyme labeling at a
concentration of 50 nM probe and 3 nM recombinant EL resulted
in time-dependent labeling of EL with
a peak signal after
120 min. To test this concept, three classes of test compounds were
tested: a known irreversible inhibitor (M-352—the parent scaffold
to the ABP); a known reversible-covalent inhibitor (M-973—a close

260
Active-site ELISA for Lipase Activity / J. Tam et al. / Anal. Biochem. 414 (2011) 254–260
Fig.7. The active-site ELISA performed under long vs short probe incubation times can identify reversible compounds. Each compound was serially diluted in DMSO and
added to 3 nM recombinant EL from HEK293 conditioned media. After 60 min ABP was added to a final concentration of 100 nM and incubated for 10 min (filled circles) or
120 min (open circles). The reaction was quenched with 1
lM M-352, and added to a streptavidin plate. The ELISA was performed using anti-EL polyclonal antibodies and
detection was by a fluorogenic peroxidase substrate as described under Materials and methods. Bars represent SEM of duplicate wells.
[5] T. Ishida, S. Choi, R.K. Kundu, K. Hirata, E.M. Rubin, A.D. Cooper, T.
Quertermous, Endothelial lipase is a major determinant of HDL level, J. Clin.
Invest. 111 (2003) 347–355.
[6] W. Jin, J.S. Millar, U. Broedl, J.M. Glick, D.J. Rader, Inhibition of endothelial
lipase causes increased HDL cholesterol levels in vivo, J. Clin. Invest. 111 (2003)
357–362.
[7] K. Ma, M. Cilingiroglu, J.D. Otvos, C.M. Ballantyne, A.J. Marian, L. Chan,
Endothelial lipase is a major genetic determinant for high-density lipoprotein
concentration, structure, and metabolism, Proc. Natl. Acad. Sci. USA 100 (2003)
2748–2753.
[8] R.J. Brown, W.R. Lagor, S. Sankaranaravanan, T. Yasuda, T. Quertermous, G.H.
Rothblat, D.J. Rader, Impact of combined deficiency of hepatic lipase and
endothelial lipase on the metabolism of both high-density lipoproteins and
apolipoprotein B-containing lipoproteins, Circ. Res. (2010).
[9] R. Gupta, P. Rathi, N. Gupta, S. Bradoo, Lipase assays for conventional and
molecular screening: an overview, Biotechnol. Appl. Biochem. 37 (2003) 63–
71.
[10] L.J. Mitnaul, J. Tian, C. Burton, M.H. Lam, Y. Zhu, S.H. Olson, J.E. Schneeweis, P.
Zuck, S. Pandit, M. Anderson, M.M. Maletic, S.T. Waddell, S.D. Wright, C.P.
Sparrow, E.G. Lund, Fluorogenic substrates for high-throughput measurements
of endothelial lipase activity, J. Lipid Res. 48 (2007) 472–482.
[11] P.M. Keller, T. Rust, D.J. Murphy, R. Matico, J.J. Trill, J.A. Krawiec, A. Jurewicz, M.
Fig.8. The ABP can be used to measure the activity of other lipase family members.
Orlistat was serially diluted in DMSO and added to 3 nM recombinant EL or LPL
from HEK293 culture media for 60 min. As a negative control, culture media from
mock-transfected cells were used. The ABP was added to a final concentration of
50 nM for 10 min before quenching with 1 lM M-352 and applying to a 96-well
streptavidin plate. Detection of bound LPL by ELISA was as described under
Materials and methods using a LPL-specific antibody. The results are representative
of two experiments, and the bars represent SEM of duplicate wells.
Jaye, M. Harpel, S. Thrall, B. Schwartz,
A high-throughput screen for
endothelial lipase using HDL as substrate, J. Biomol. Screen. 13 (2008) 468–
475.
[12] W. McNeely, P. Benfield, Orlistat. Drugs 56 (1998) 241–249.
[13] R. Guerciolini, Mode of action of orlistat, Int. J. Obes. 21 (1997) S12–S23.
[14] M.L. Drent, I. Larsson, T. Williamolsson, F. Quaade, F. Czubayko, K.
Vonbergmann, W. Strobel, L. Sjostrom, E.A. Vanderveen, Orlistat (Ro-18–
0647), a lipase inhibitor, in the treatment of human obesity—a multiple-dose
study, Int. J. Obes. 19 (1995) 221–226.
[15] A.M. Gold, D.E. Fahrney, The mechanism of reactivation of
phenylmethanesulfonyl alpha-chymotrypsin, Biochem. Biophys. Res.
Commun. 10 (1963) 55–59.
Acknowledgment
The authors thank Emanuela Nizi for synthesizing the probe
and Louis Jacques Fortin for compound management assistance.
[16] W. Jin, G.S. Sun, D. Marchadier, E. Octtaviani, J.M. Glick, D.J. Rader, Endothelial
cells secrete triglyceride lipase and phospholipase activities in response to
cytokines as a result of endothelial lipase, Circ. Res. 92 (2003) 644–650.
[17] K. Hirata, T. Ishida, H. Matsushita, P.S. Tsao, T. Quertermous, Regulated
expression of endothelial cell-derived lipase, Biochem. Biophys. Res. Commun.
272 (2000) 90–93.
[18] U.C. Broedl, C. Maugeais, D. Marchadier, J.M. Glick, D.J. Rader, Effects of
nonlipolytic ligand function of endothelial lipase on high density lipoprotein
metabolism in vivo, J. Biol. Chem. 278 (2003) 40688–40693.
[19] W. Jin, I.V. Fuki, N.G. Seidah, S. Benjannet, J.M. Glick, D.J. Rader, Proprotein
convertases [corrected] are responsible for proteolysis and inactivation of
endothelial lipase, J. Biol. Chem. 280 (2005) 36551–36559.
[20] M. Gauster, A. Hrzenjak, K. Schick, S. Frank, Endothelial lipase is inactivated
upon cleavage by the members of the proprotein convertase family, J. Lipid
Res. 46 (2005) 977–987.
References
[1] K. Hirata, H.L. Dichek, J.A. Cioffi, S.Y. Choi, N.J. Leeper, L. Quintana, G.S. Kronmal,
A.D. Cooper, T. Quertermous, Cloning of a unique lipase from endothelial cells
extends the lipase gene family, J. Biol. Chem. 274 (1999) 14170–14175.
[2] M.G. McCoy, G.S. Sun, D. Marchadier, C. Maugeais, J.M. Glick, D.J. Rader,
Characterization of the lipolytic activity of endothelial lipase, J. Lipid Res. 43
(2002) 921–929.
[3] M. Jaye, K.J. Lynch, J. Krawiec, D. Marchadier, C. Maugeais, K. Doan, V. South, D.
Amin, M. Perrone, D.J. Rader, A novel endothelial-derived lipase that modulates
HDL metabolism, Nat. Genet. 21 (1999) 424–428.
[4] C. Maugeais, U.J. Tietge, U.C. Broedl, D. Marchadier, W. Cain, M.G. McCoy, S.
Lund-Katz, J.M. Glick, D.J. Rader, Dose-dependent acceleration of high-density
lipoprotein catabolism by endothelial lipase, Circulation 108 (2003) 2121–
2126.
[21] S.L. McGovern, B.T. Helfand, B. Feng, B.K. Shoichet, A specific mechanism of
nonspecific inhibition, J. Med. Chem. 46 (2003) 4265–4272.